处理 CAPTCHA 事件

了解 无头浏览器 如何捕获和处理 CAPTCHA 事件。

基于事件的方法

要监控 CAPTCHA 求解过程，我们建议使用基于事件的方法。它允许你的脚本监听直接由扩展发送的自定义消息。

工作原理

该扩展向浏览器的 window 对象发送消息。你可以通过添加一个“message”事件监听器来捕获这些消息。

oxylabs-captcha-solve-start	检测到 CAPTCHA，且求解已开始。
oxylabs-captcha-solve-end	CAPTCHA 已成功解决。
oxylabs-captcha-solve-error	自动求解器未能绕过 CAPTCHA。

你可以订阅这些事件，在 CAPTCHA 成功处理后暂停自动化并恢复。通常，CAPTCHA 求解最多需要 30 秒，具体取决于 CAPTCHA 的类型和复杂度，但在某些情况下可能更久，因此请预留足够时间等待 CAPTCHA 求解结束。

代码示例

Python（Playwright）

import sys
from playwright.sync_api import sync_playwright

# Configuration
EXTENSION_NAME = 'oxylabs-runtime'
CAPTCHA_SOLVE_END = 'oxylabs-captcha-solve-end'
CAPTCHA_SOLVE_ERROR = 'oxylabs-captcha-solve-error'
UB_BROWSER_URL = 'wss://<username>:<password>@ubc.oxylabs.io'
CAPTCHA_CHECK_TIMEOUT_SEC = 60
TARGET_URL = 'https://www.indeed.com/cmp/Bank-of-the-West/reviews?lang=any&fcountry=ALL&sort=date'

def run():
    with sync_playwright() as p:
        print('Opening browser page...')
        # Connect to the remote browser using CDP
        browser = p.chromium.connect_over_cdp(UB_BROWSER_URL)
        
        # Get the default context and create a new page
        ctx = browser.contexts[0]
        page = ctx.new_page()

        # This must be done BEFORE navigation.
        ctx.add_init_script(f"""
            window.addEventListener("message", (event) => {{
                if (event.data && event.data.source === "{EXTENSION_NAME}") {{
                    window.__extensionStatus = event.data.type;
                }}
            }});
        """)

        print('Opening target website...')
        page.goto(TARGET_URL, wait_until='domcontentloaded')

        try:
            # Wait for the extension status to change
            page.wait_for_function(
                """
                ([solveEnd, solveError]) => {
                    const status = window.__extensionStatus;
                    if (status === solveError) {
                        throw new Error("CAPTCHA solving failed");
                    }
                    return status === solveEnd;
                }
                """,
                arg=[CAPTCHA_SOLVE_END, CAPTCHA_SOLVE_ERROR],
                timeout=CAPTCHA_CHECK_TIMEOUT_SEC * 1000
            )
            print('CAPTCHA solved successfully, continue scraping...')
            
        except Exception as err:
            print(f'Error during CAPTCHA solve: {err}')
            browser.close()
            sys.exit(1)

        page.wait_for_selector('h1[data-testid="PageHeader-title-reviews"]')
        page.screenshot(path='page_screenshot.jpg')

        page.close()
        browser.close()
        print('Done.')

if __name__ == "__main__":
    run()

JavaScript（Playwright）

JavaScript（Puppeteer）